Cucumis melo L. var. cantalupensis Naudin - American Chemical

Chapter 4

Downloaded by CORNELL UNIV on May 31, 2012 | http://pubs.acs.org Publication Date (Web): May 29, 2012 | doi: 10.1021/bk-2012-1098.ch004

Confirmation of Trace Level AromaImpact Compounds in Cantaloupe (Cucumis melo L. var. cantalupensis Naudin) by GC-MS/MS Analysis J. Lin* and Y. Wang Firmenich Inc., North America R&D, P.O. Box 5880, Princeton, NJ 08543, U.S.A. *E-mail: [email protected]

A room temperature sample preparation method combining solid phase extraction (SPE) and gel permeation chromatography (GPC) was used to obtain an authentic aroma extract of fresh cantaloupe purchased from a local farm. More than thirty aroma-impact compounds of the cantaloupe were determined by GC-O analysis. No EI mass spectra were obtained for the unequivocal identification of ten trace level aroma-impact compounds. To demonstrate the usefulness of GC-MS/MS analysis in the identification of trace level aroma-impact compounds, selected reaction monitoring (SRM) GC-MS/MS methods were developed for simultaneous analysis of six C9 compounds by choosing GC columns and MS ionization modes, optimization of the CI source temperature and pressure, and collision energy, and selection of product ions. The application of SRM GC-MS/MS analysis led to the confirmation of 6Z-nonenal, 2E-nonenal and 2E,6Z-nonadienal and the unambiguous identification of 3Z-nonenyl acetate and 3Z,6Z-nonadienyl acetate in cantaloupe. The presence of 3Z,6Z-nonadienal in cantaloupe was not confirmed by our GC-MS/MS analysis.

© 2012 American Chemical Society In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Introduction The inherent sensitivity and selectivity of gas chromatography-tandem mass spectrometry (GC-MS/MS) make it a powerful tool for challenging food product analysis. The main application of GC-MS/MS is for quantification of trace level analytes in complex matrices, such as mycotoxins and contaminants in foods (1, 2), and pesticide residues in agricultural products (3, 4). GC-MS/MS has also been used in qualitative targeted compounds analysis. The presence of khusimone (5) and, khusimene and acoradiene (6) in vetiver oil were confirmed in very sensitive and selective ways using combined GC-MS/MS analysis, i.e. daughter ion scan with neutral loss scan and, parent ion scan with neutral loss scan, respectively. Recently, 2E,6Z-nonadienyl acetate was identified for the first time in melon using a sensitive and selective selected ion monitoring GC-MS/MS method (7). In the aroma extract dilution analysis of muskmelon, 46 aroma impact compounds were detected (8). Only 15 of them were identified on the basis of mass spectra, retention indices and odor descriptors. The others remained unknown, most likely because these odorants occurred at trace levels and couldn’t be identified by GC-MS. Identification of these unknown odorants can be achieved by proposing molecules based on aroma characteristics, retention index, as well as other clues. GC-MS/MS methods can then be used to confirm or disapprove the initial structural proposals. In this work, GC-O analysis was conducted on an aroma extract of cantaloupe to determine aroma impact compounds. Sensitive GC-MS/MS methods were then developed to confirm the presence of several trace level C9 aroma-impact compounds in cantaloupe. Our goal was to use this as an example to demonstrate the value of sensitive and selective GC-MS/MS analyses in the identication of trace level aroma impact compounds.

Experimental Section Plant Materials Fresh ripe cantaloupes (Cucumis melo L. var. cantalupensis Naudin) were purchased from a local farm near Princeton, NJ. The mesocarp tissues of a 2.7 kg cantaloupe were cut into 1 inch cubes and immediately comminuted to puree with a Waring Juice Extractor. The puree was then centrifuged at 15°C at a speed of 8000 rpm for 10 min. About 1 L of the supernatant was obtained and used for aroma extraction by SPE described below. No effort was made to inactivate enzymes in order to be consistent with how melon is commonly eaten. The mesocarp tissues of the other melons were also cut into 1 inch cubes and then stored under vacumn at -15 °C until use.

42 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.

Preparation of Cantaloupe Aroma Extract


Solid Phase Extraction (SPE)

Three Oasis® HLB cartridge (20cc/1g, 60 μm) (Waters, Milford, MA) were all conditioned sequentially with 10 mL dichloromethane, 20 mL methanol and 30 mL 18 MΩ water. The juice supernatant (1 L total, 500 mL each) was passed through two conditioned cartridges at a flow rate of approximately 8 mL/min. The sample load was collected and combined and passed through the third cartridge. Each cartridge was washed with 60 mL of water then eluted with 17 mL dichloromethane/methanol (9/1, v/v). The eluates from the three cartridges were combined and dried with anhydrous magnesium sulfate and then filtered. The filtrate was concentrated to a final volume of about 12 mL with a gentle stream of nitrogen and then evaluated organoleptically on a perfume blotter. Ten mL of the extract was further concentrated to about 2.5 mL with a gentle stream of nitrogen and some precipitation was observed during this process.

Gel Permeation Chromatography (GPC)

This was performed on a GPC system modified from a Biotage SP1™ Purification System, which was equipped with a pump, a UV detector and an automatic fraction collector. A manual sample injector (model 7725, with a 5mL sample loop, Rheodyne) was connected to the system between the exit of the pump and the head of column for sample introduction. One three-way valve (Omnifrit, purchased from VWR) was connected to the end of the column for ease of operation. With such a valve, the column flow can be switched to the fraction collector, waste, or simply stopped. The column (25 mm i.d. x 500 mm l., with 25 μm PTFE frits., Omnifrit) was packed with 50 g of Biobeads® S-X3 using 150 mL of DCM according to the directions provided by Bio-RAD. The bed length of the column was 400 mm. Sample (2.5 mL) obtained from SPE was manually injected. DCM was used as the mobile phase. The flow rate was 5 mL/min. The eluent was collected into 6 mL fractions and the fractions were evaluated with perfume blotters. Fractions 15 to 24 were found to have aroma and therefore combined and concentrated to a final volume of 0.5 mL. This aroma extract was then analyzed by GC-O/MS and GC-MS/MS.



GC-O/MS Analysis This was performed on a custom-designed Multidimensional AromaTrax™ GC-O System (Microanalytics, TX). It integrates an Agilent 6890 based dual column multidimensional gas chromatograph (MDGC) with an Agilent model 5973N mass spectrometer and an olfactory port with humidifier. The MDGC is installed with dual columns, including Rxi-1MS column (30m x 0.53 mm i.d. x 0.5 μm, Restek) as the first column and DB-Wax (30 m x 0.53 mm i.d. x 0.5 μm, J&W) as the second column. Samples (1 μL) for GC-O analysis were directly introduced to the second column via splitless injection. The eluent of the second column was split between the MS and the olfactory detectors at a ratio of about 1/5 via an open split interface. Six GC-O runs with three experienced panelists were used to determine aroma-impact compounds.

GC-MS/MS Analysis GC-MS/MS experiments were performed on a Varian 3800 GC coupled to a Varian 300 MS. The system was equipped with a CombiPal autosampler, two split/splitless injectors, a DB-WaxETR column (30m x 0.25 mm id x 0.25 μm, J&W) on the front injector and a DB-17MS column (20m x 0.18 mm id x 0.18 μm, J&W) installed on the back injector. The ends of both columns combine and transfer to the MS via a zero-dead-volume T-union (Valco) and a restrictor (~0.3 m x 0.11mm i.d.).

Synthesis of Standards Standard compounds were needed to develop GC-MS/MS methods. 3Z,6ZNonadienyl acetate, 3Z-nonenyl acetate and 3Z,6Z-nonadienal were not available commercially or in house. They were prepared according to the synthesis schemes shown in Figure 1. The targets obtained were characterized by NMR.

Figure 1. Synthesis schemes for the standard compounds. 44 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.

Results and Discussion


Isolation of Volatile Compounds Initially, we planned to isolate the volatile compounds of cantaloupe via hydrodistillation under vacuum, followed by solid phase extraction (SPE). The aroma of the hydrodistillate obtained was described as cooked, meaty, sulfury, slightly melon, although the same method applied to honeydew melon resulted in an aroma extract resembling that of the original fruit (7). A room temperature volatile isolation method combining SPE and gel permeation chromatography (GPC) cleanup was then devised and applied, giving rise to a very authentic cantaloupe aroma extract. GPC was effective in removing co-extracted carotenoids, waxes and other lipophilic components because their molecular sizes are larger than those of aroma molecules. The main advantage of this volatile isolation method is the absence of thermal stress. Additionally, it is quick and easy to perform. Aroma-Impact Compounds Based on GC-O analysis, 33 odorants were determined to be aroma-impact compounds of the cantaloupe sample (Table 1). The aromagram plotting these 33 compounds versus their total GC-O intensity is inversely overlaid with the GC-MS chromatogram of the cantaloupe aroma extract in Figure 2. The six most intense aroma-impact compounds were 3Z-hexenyl acetate (11, fruity), 3Z-hexen1-ol (13, green), ethyl 2-(methylthio)acetate (16, potato, fruity), 6E-nonen-1-ol (24, melon, sweet), 3Z,6Z-nonadien-1-ol (27, fresh, melon) and phenethanol (29, sweet). The odor characteristics of these compounds were in good agreement with the overall cantaloupe aroma impression described as melon, fresh, fruity and sweet. Sixteen out of the 33 aroma-impact compounds were lipid degradation products, including saturated and unsaturated fatty alcohols, aldehdyes and ketones. The alcohols had higher odor thresholds therefore, they must be present at higher concentrations to contribute to the overall aroma. Their mass spectra were acquired for their identification. However, unsaturated aldehydes and ketones are often very potent odorants. It is extremely difficult to obtain their EI spectra to confirm their identities. As shown in Table 1, 3Z-hexenal (8), 1,5Z-octadien-3-one (12), 6Z-nonenal (15), 2Z-nonenal (18), 2E-nonenal (19), 2E,6E-nonadienal (20), 2E,6Z-nonadienal (21)and trans-4,5-epoxy-2E-decenal (31) were identified based on aroma characteristics and retention indices on two columns of different polarity. This is commonly realized via an experienced GC-O panelist’s recognizing the characteristic odor at the GC sniff port, followed by further confirmation by RIs on two different columns. In some cases, due to overlapping with another odorant, a panelist might not be able to recognize the odor characteristics on another column, and therefore it may not be possible to confirm the identification by this means. On the other hand, library and literature data on retention indices and aroma descriptors can be a good starting point to propose candidate molecules for the identification of trace aroma-impact compounds. Other evidence, such as quantitative structure-retention relationship 45 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.

and the presence of other homologues, isomers, precursors or degradation products, can be used to propose structures for trace level unknown aroma-impact compounds. Inexperience in recognizing odor characteristics may lead to the inability to identify trace aroma-impact compounds or misidentification. Therefore, sensitive and selective methods that do not depend on olfactive acumen and experience can be very useful to confirm the structure proposals of trace level aroma-impact compounds.


Table 1. Aroma-impact compounds identified in the cantaloupe aroma extract

No.

ID

RI* (GC-O)

RI (inj/ lit.)

Descriptors

Total Int.

ID criteria

1

propyl acetate

985

986

solventy, estery, pineapple, fruity

174

RI, MS, O

2

methyl butyrate

995

995

butric, cheesy, sharp

159

RI, MS, O

241

RI, MS, O

3

ethyl acrylate

999

1000

pungent, plastic, fruity, slightly tropical

4

ethyl butyrate

1047

1046

cheesy, estery, pineapple, fruity

168

RI, MS, O

5

methyl thioacetate

1054

sulfury, mango, sulfury-pineapple, tropical fruit

231

RI, MS, O

6

ethyl 2-methylbutyrate

1061

1060

slightly fruity, green, apple, sweet, honey

191

RI, MS, O

7

hexanal

1085

1092

green, leafy

150

RI, MS, O

169

RI, RI, O

1055

8

3Z-hexenal

1144

1148

green, leafy, hexanol, melon, aldehydic

9

eucalyptol

1209

1202

camphoraceous, menthol

179

RI, MS, O

10

2-methyl-1butanol

1212

1212

chocolate, green, musty, cheesey

180

RI, MS, O

11

3Z-hexenyl acetate

1320

fruity, olive, creamy, green apple, pea

386

RI, MS, O

1380

geranium, vegetative, metallic, bloody

167

RI, RI, O

1321

1,5Z-octadien-3one 12

1375

Continued on next page.


Table 1. (Continued). Aroma-impact compounds identified in the cantaloupe aroma extract

No.

13


14

ID

3Z-hexen-1-ol

unknown

15

6Z-nonenal

16

ethyl 2-(methylthio)acetate

RI* (GC-O)

1390

1433

1444

1451

RI (inj/ lit.)

Descriptors

Total Int.

ID criteria

1396

fruity, herbaceous, leafy, green, aldehyde

331

RI, MS, O

NA

sulfury, earthy, musty, cardboard, dirty

262

NA

1442

creamy, sweet, watermelon, dairy, lactonic

280

RI, RI, O

1452

potato, fruity, green, vegetative, tomato, aldehyde

445

RI, MS, O

1497

mushroom, bloody, metallic, green

219

RI, RI, O

159

RI, RI, O

1,5Z-octadien3-ol 1492

17

18

2Z-nonenal

1506

1509

geranium, aldehyde, fruity, floral

19

2E-nonenal

1536

1538

soapy, chalky, waxy, cardboard

255

RI, RI, O

1566

geranium, aldehyde, gamma lactone, candy, metalic

216

RI, RI, O

1578

fresh cucumber, green pepper, green, rindy

246

RI, RI, O

252

RI, MS, O

20

2E,6Enonadienal

21

2E,6Znonadienal

22

3-(methylthio)propyl acetate 1629

1618

potato, mushroom, metallic, soysauce, meaty, fried

23

1-nonanol

1666

fruity, geranium, green, soapy

189

RI, MS, O

330

RI, MS, O

185

RI, MS, O

1566

1583

1668

24

6E-nonen-1-ol

1712

1713

sweet, melon, creamy, candy, lactonic, dairy

25

methionol

1719

1722

vegetative, green pepper, rubbery

Continued on next page.


Table 1. (Continued). Aroma-impact compounds identified in the cantaloupe aroma extract


No.

26

6Z-nonen-1-ol

27

3Z,6Z-nonadien1-ol

28

2E,6Z-nonadien1-ol

29

phenethanol

1720

1756

1773

1906

Total Int.

ID criteria

1721

fresh, seaweed, green, cucumber, fatty, melon

219

RI, MS, O

1758

fresh, melon, aldehydic, cucumber, green

382

RI, MS, O

1770

fresh, cucumber, melon, aldehyde, green, candy

213

RI, MS, O

1916

sweet, hay, cinnamon, creamy, lactone, musky

334

RI, MS, O

1937

paprika, oakmoss, algae like, green,metallic, sweet

214

RI, MS, O

253

RI, RI, O

Descriptors

beta-ionone

31

trans-4,5-epoxy2E-decenal

2002

2000

green,metallic, rusty, fruity, sweet, hay, weak

32

Furaneol®**

2038

2039

caramel, maltol, hay-like, sweet

249

RI, RI, O

NA

sweet, spicy, dried fruit, woody, cinnamon

174

NA

unknown

RI on a DB-Wax column;

1936

RI (inj/ lit.)

30

33 *

ID

RI* (GC-O)

2077 **

Registered trademark of Firmenich SA, Switzerland

GC-MS/MS Analysis of C9 Aroma-Impact Compounds Ten C9 compounds have been revealed by GC-O analysis to be important aroma contributors of cantaloupe as shown in Table 1, but five of them (15, 18, 19, 20 and 21) occurred in the extract at trace levels. It was not possible to confirm their identification by MS. In addition, 3Z-nonenyl acetate had been identified and 3Z,6Z-nonadienyl acetate tentatively identified in cantaloupe by Kourkoutas et al. (9). 3Z,6Z-Nonadienal is a degradation product of linolenic acid and was thought to be the precursor of 3Z,6Z-nonadien-3-ol and 2E, 6Z-nonadienal (10). These three compounds were not detected in our cantaloupe aroma extract by GC-MS analysis. The structures of the eight C9 compounds are shown in Figure 3. Next, the development of GC-MS/MS method for the detection of these C9 compounds in cantaloupe will be presented. 48 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Figure 2. Aromagram of the cantaloupe extract as compared to its GC-MS chromatogram.

Optimization of GC and MS/MS Conditions Stock solutions containing six target compounds (15, 19, 21, 34, 35 and 36) were prepared and used to optimize GC-MS/MS conditions in terms of resolution, ionization modes, CI source temperature and pressure, collision energy and transition ions. All the six compounds were well resolved from each other on a DB-WaxETR column. However, severe peak broadening of 3Z,6Z-nonadienal was observed, most likely due to its isomerization to 2E,6Z-nonadienal on the polar stationary phase. Fortunately, 3Z,6Z-nonadienal eluted as a sharp peak on a Rxi-1ms column. However, 3Z,6Z-nonadienyl acetate and 3Z-nonenyl acetate were not resolved on this column. A DB-17MS column was found to provide excellent peak shapes and satisfactory resolution of all the compounds although 6Z-nonenal and 3Z,6Z-nonadienal were not baseline resolved. 49 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


Figure 3. Chemical strucrures of C9 aroma compounds found or suspected in cantaloupe.

The mass spectra, i.e. EI-MS, PCI (CH4)-MS, NCI (CH4)-MS, PCI (NH3)-MS and NCI (NH3)-MS, of the standard compounds were examined to select the best ionization mode. The principle of this selection process is to choose an ionization mode that gives strong molecular ions and the lowest detection limits for the benefit of sensitivity and selectivity. The EI spectra did not have the expected molecular ions and there were no other good candidate precursor ions for MS/MS experiments. No intense molecular ions were observed in the PCI (CH4) spectra due to further fragamentation into small ions. The sensitivity of PCI with CH4 as the reagent gas was found to be best at a source temperature of 180 °C and a source pressure of 8 torr. No good NCI (CH4) spectra of the standard compounds were obtained. When using NH3 as the reagent gas, intense molecular ions of all the standard compounds appeared in their spectra in both positive and negative modes. The sensitivity of PCI and NCI with NH3 as reagent gas was optimized at a source temperature of 150 °C and a source pressure of 6 torr. NCI was found to be much more sensitive than PCI for all the standard compounds. Therefore, NCI (NH3) was selected as the best ionization mode for this study. Product ion spectra of the standard compounds’ molecular ions [M-1]were acquired at various collision energy (CE) levels (4, 8, 12 and 16 eV) and investigated. This led to the selection of two product ions of each compound for selected reaction monitoring (Table 2). The principle that guided this selection process was to choose the product ions giving the best sensitivity and specificity. For the benefit of sensitivity, product ions should be strong in intensity and have large intensity increases when the intensities of the precursor ions were reduced by varying the CE level. In general, product ions with higher m/z values have better specificity. Due to the nature of these compounds, the selected product ions are not of high specificity. If an interference is encountered in a sample analysis, 50 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


it might be overcome by choosing a different product ion. As an example, the product ion spectrum of ion m/z 137 at 12.0 eV for 2E,6Z-nonadienal is shown in Figure 4. An intense peak at m/z 69, which could be attributed to the neutral loss of 1,3-pentadiene, was observed and selected as one product ion. Ion m/z 119, likely produced by a neutral loss of water from ion m/z 137, was chosen as the second product ion.

Figure 4. Product ion spectrum of ion m/z 137 of 2E,6Z-nonadienal at a collision energy of 12.0 eV, with structural explanation for transitions m/z 137→69 and m/z 137→119.

Once the transitions to be monitored had been selected for each compound, the CE was further optimized to obtain the most intense signals for both transitions. In practice, this was conducted by varying the CE by large increments (4 eV), followed by fine-tuning of the CE (1 eV increments). Comparing the relative intensities of the two transitions of each compound at different CEs, it was straight forward to find the the best CE for each transition as shown in Table 3. When the best CEs for the two transitions were different, we decided to sacrifice the intensity of the more intense transitions and optimize the CE to favor the weaker transitions. Taking compound 2E-nonenal as an example, according to the data shown in Table 3, 8 eV and 12 eV were the best CEs for transitions m/z 139→69 and m/z 139→82 respectively. The CE was optimized at 12 eV to favor the weaker transition m/z 139→82. After fine-tuning the CE, the optimum CEs were obtained as listed in Table 2. 51 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.

Table 2. The SRMs and the optimum collision energy (CE) of the standard compounds MW

Precursor ion (m/z)

6Z-nonenal

140

139

2E-nonenal

140

139

2E,6Z-nonadienal

138

137


Compound Name

3Z-nonenyl acetate

Product ion (m/z) 57 83 69 82 69 119 153

184

183

3Z,6Z-nonadienylacetate

182

181

3Z,6Z-nonadienal

138

137

41 59 151 69 119

Collision Energy (eV) 12

12

8

8

8

8

Identification of C9 Aroma Compounds in Cantaloupe by GC-MS/MS Analysis With the optimized GC and MS/MS conditions, two GC-MS/MS methods, one using the DB-WaxETD column and the other using the DB-17MS column, were set up and applied to identify or confirm the targeted C9 compounds in cantaloupe. Three samples, cantaloupe aroma extract, the aroma extract spiked with standards and a solution of standard compounds, were prepared and analyzed using both methods. The samples were analyzed in the above order with a blank sample before and after each sample. The relative intensities of any two SRMs of a given compound under particular conditions is specific to the compound. Therefore it is considered as one Identification Point (IP) (11). The accepted deviation in relative intensity depends on the scale of the relative intensity . The relative intensities of the two SRMs of each compound were calculated and compared among the three samples. The results are summarized in Table 4 along with the retention times and the accepted deviation in relative intensity. As shown in Table 4, the relative intensities of the two SRMs for each compound were extremely consistent among the three samples, with the exception of 2E-nonenal on the DB-WaxETR column. These consistencies in relative intensity (1 IP), one precusor ion (1 IP) and two product ions (2 IPs) added to a total of 4 IPs for the positive identification of the compounds. Therefore, the GC-MS/MS analysis with the DB-WaxETR column led to the confirmation 52 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.


of trace level aroma-impact compounds 6Z-nonenal and 2E,6Z-nonadienal in cantaloupe. Furthermore, 3Z-nonenyl acetate and 3Z,6Z-nonadienyl acetate were unambiguously identified in cantaloupe in this analysis. The presence of 2E-nonenal in cantaloupe was confirmed only by the GC-MS/MS analysis with the DB-17MS column because an interference was encountered in the analysis with the DB-WaxETR column. Finally, this analysis did not lead to the identification of 3Z,6Z-nonadienal in cantaloupe.

Table 3. Effect of collision energy (CE) on signal intensities of each SRM Compound Name

Transitions

Relative signal intensity

CE=8 eV

CE=12 eV

CE=16 eV

139→57

10

32

40

35

139→83

60

170

75

30

139→69

100

250

190

100

139→82

5

60

120

75

137→69

90

400

310

190

137→119

60

125

60

5

3Z-nonenyl acetate

183→153

1.0

1.0

0.5

0.35

183→41

0.45

1.15

0.75

0.25

3Z,6Znonadienylacetate

181→59

2.7

2.7

2.3

1.5

181→151

75

190

180

60

137→69

0.25

1.4

1.6

1.1

137→119

1.65

2.5

0.8

0.3

CE=4 eV 6Z-nonenal

2E-nonenal 2E,6Znonadienal

3Z,6Znonadienal

GC-MS/MS chromatograms of SRMs m/z 181→59 and m/z 181→151 in the standard 3Z,6Z-nonadienyl acetate solution, cantaloupe aroma extract and the aroma extract spiked with the standards were compared and shown in Figure 5 as an example. This figure shows the excellent selectivity of the method because no other surrounding interfering peaks obscured the identification of the compound. The high signal-to-noise ratios of the peaks were due to the sensitive and selective nature of GC-MS/MS method. This in turn allowed accurate determination of the peak areas and increased the reliability of the calculated relative intensities which contributed to the consistency of the relative intensities shown in Table 4. 53 In Recent Advances in the Analysis of Food and Flavors; Toth, S., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2012.

Table 4. Retention times and relative intensities of the SRMs of each compound, compared to the cantaloupe aroma extract and the spiked extract Compound

Sample

Standard compounds

Aroma extract

spiked aroma extract

Accepted deviation*

Analyzed on the DB-WaxETR column 6Z-nonenal

2E-nonenal Downloaded by CORNELL UNIV on May 31, 2012 | http://pubs.acs.org Publication Date (Web): May 29, 2012 | doi: 10.1021/bk-2012-1098.ch004

2E,6Znonadienal 3Z-nonenyl acetate 3Z,6Znonadienyl acetate

RT (min)

15.34

15.33

15.36

—

Rel. Int.(%)

18

17

17

5

RT (min)

17.40

17.40

17.41

—

Rel. Int.(%)

74

37

48

15

RT (min)

18.58

18.59

18.60

—

Rel. Int.(%)

34

33

37

9

RT (min)

19.09

19.01

19.13

—

Rel. Int.(%)

68

70

71

14

RT (min)

20.45

20.47

20.51

—

Rel. Int.(%)

3

3

3

2

Analyzed on the DB-17MS column 6Z-nonenal

2E-nonenal 2E,6Znonadienal 3Z-nonenyl acetate 3Z,6Znonadienyl acetate 3Z,6Znonadienal *

RT (min)

11.52

11.61

11.63

—

Rel. Int.(%)

12

12

13

4

RT (min)

12.98

13.05

13.03

—

Rel. Int.(%)

74

68

83

15

RT (min)

13.34

13.41

13.38

—

Rel. Int.(%)

41

41

45

10

RT (min)

15.98

16.14

16.14

—

Rel. Int.(%)

76

71

76

15

RT (min)

16.50

16.67

16.67

—

Rel. int.(%)

3

3

4

2

RT (min)

11.60

nd

11.65

—

Rel. Int.(%)

13

nd

18

4

Based on the Identification Points (IPs) recommended in the European directive (11).



Figure 5. SRM (m/z 181→59 and m/z 181→151) signals of 3Z,6Z-nonadienyl acetate in standard solution, cantaloupe aroma extract and the aroma extract spiked with the standard.

Conclusions Confirmation of trace level aroma-impact compounds in complex aroma extracts can be achieved by a combination of GC-MS/MS analyses. The application of sensitive and selective GC-MS/MS methods to cantaloupe aroma extract led to the confirmation of 6Z-nonenal, 2E-nonenal and 2E,6Z-nonadienal in cantaloupe and the unambiguous identification of 3Z-nonenyl acetate and 3Z,6Z-nonadienyl acetate in cantaloupe. The presence of 3Z,6Z-nonadienal in cantaloupe was not confirmed by our GC-MS/MS analysis. It would be interesting to explore the deployment of other sensitive and selective GC-MS/MS methods in various flavor related studies.


Acknowledgments We would like to thank botanist Dr. Yong Ming Yuan for his assistance in the classification of melon, and flavorists Jung-A Kim and Fred Triolo, Jr. for their valuable help in GC-sniffing..

References Melchert, H.-U.; Pabel, E. J. Chromatogr. A 2004, 1056 (1-2), 195–199. Abdulkadar, A. H. W.; Kunhi, A. A. M.; Jassim, A.-J.; Abdulla, A.-A. Food Addit. Contam. 2003, 20 (12), 1164–1169. 3. Lehotay, S. J. J. AOAC Int. 2000, 83 (3), 680–697. 4. Cervera, M. I.; Medina, C.; Portoles, T.; Pitarch, E.; Beltran, J.; Serrahima, E.; Pineda, L.; Munoz, G.; Centrich, F.; Hernandez, F. Anal. Bioanal Chem. 2010, 397 (7), 2873–2891. 5. Cazaussus, A.; Pes, R.; Sellier, N.; Tabet, J. C. Chromatographia 1988, 25 (10), 865–869. 6. Sellier, N.; Cazaussus, A. J. Chromatogr. 1991, 557, 451–458. 7. Perry, P. L.; Wang, Y.; Lin, J. Flavour Fragrance J. 2009, 24, 341–347. 8. Hayata, Y.; Sakamoto, T.; Maneerat, C.; Li, X.; Kozuka, H.; Sakamoto, K. J. Agric. Food. Chem. 2003, 51, 3415–3418. 9. Kourkoutas, D.; Elmore, J. S.; Mottram, D. S. Food Chem. 2006, 97, 95–102. 10. Hatanaka, A.; Kajiwara, T.; Harada, T. Phytochemistry 1975, 14 (12), 2589–2592. 11. Commission decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Off. J. Eur. Communities 2002, L221 (2002/657/EC), 8–36.


1. 2.


Cucumis melo L. var. cantalupensis Naudin - American Chemical

Recommend Documents